@DavidCarlisle -- urk. but not unexpected, at least by me. sigh. i can't do anything useful from this laptop, but when i'm at my workstation tomorrow, i'll send you the search argument i used. but i'm sure it has only spaces in it.

@AlanMunn Thank God I don't know what Marmite is. I think it's the thing a colleague of mine (who worked in the US) once tried and told us it tasted like a dead rat. BTW, we were talking about "the worst thing you've ever eat" after that a Serbian colleague of ours offer us some (disgusting) banana biscuits dressed by chocolates she bring us from her home country.

@AlanMunn Funny anecdote regarding the same Serbian colleague: once she went to a bakery to buy some bread but instead of saying "vorrei del pane" (= I'd like some bread) she said "vorrei il pene" (= I'd like the penis) and she was wondering why the baker was ROFL.

@barbarabeeton yes I've always known the sql is only looking for the four-space versions, but while it was still showing thousands of posts to edit, the fact that it missed the tab ones wasn't really a concern. Perhaps @Moriambar can find a variant that works, or we could go back to the powers and demand one. If I remember correctly the site search is effectively looking for \ then using table and other tags to make a real control space less likely.

@UlrikeFischer Fixed that issue but there is one problem in pst-pdf: it uses \GPT@page and has to have it defined. I'll ping the author and ask for an update

@DavidCarlisle AFAIK the tab is translated into spaces. What my previous query missed was about post that were not accepted but had a null accepted answer id. The incriminated post was found by my query regarfless of spaces. Also the "tab" query gave false positives. Unfortunately I'm stuck at work and my boss really wants me to do what I'm paid for (ie the monkey) so...

@DavidCarlisle @Moriambar I lost track of all this queries ;-). I thought I had handled all of mine. But yesterday @barbarabeeton edited one of my answers which I had missed. Is there some actual "state of the art" query to search by user id?

@UlrikeFischer nope. We're working on year based queries

@JosephWright If necessary I can reach Rolf. There is a xetex/pstricks question on the main site. I'm just looking (have to update first).

@UlrikeFischer I've put in a temporary fix for the \GPT@page business: the auto-pst-pdf example then compiles

@DavidCarlisle @barbarabeeton anyway I think the best procedure would be to process my "old query" first, the year based one. Those have to be processed anyway and does not intersect with the new one, which I will fix as soon as possible

@Moriambar going to work now but nearly done jabuary 2013, just 4 or 5 left

@DavidCarlisle great, but don't do too many, give me time to fix the query

:P

@UlrikeFischer I've dropped him a line

@JosephWright the pstricks question looks more serious. As far as I can see the color model (at least for gray) in xetex doesn't fit the expection of pstricks/\c@lor@to@ps. The log shows \color@black ->\xcolor@ {}{[0]}{gray}{0} instead of \color@black ->\xcolor@ {}{gray 0}{gray}{0} and due to the missing space after gray \c@lor@to@ps fails.

@UlrikeFischer Already fixed and sent to CTAN

@JosephWright OK: I will add a comment to the question.

@UlrikeFischer Link? I can post an answer

@JosephWright I always forget the links. tex.stackexchange.com/questions/374863/…

@JosephWright Is there a work-around? miktex is sometimes a bit slower to include such updates.

Can any LaTeX engine use WOFF2 fonts?

Apparently, our company is changing looks again.

@UlrikeFischer Basically copy all of the colour part of dvips.def and stick it into the preamble: I've not worked out a version of \c@lor@to@ps that would work with the dvipdfmx case

@wilx hmmm woff are not friends of pdfs. They are meant for the internet not for embedding

@JosephWright Yes, I just found the commit where you did all the changes. Getting xetex.def from CTAN is easier ;-)

@UlrikeFischer Indeed

@UlrikeFischer I'll post an answer anyway

I'll try this: google/woff2

Nope. Windows Font Viewer does not like the resulting file.

@DavidCarlisle I have good news. The "other query", ie the one which prompted around 2000 more posts, is not to be trusted since includes many already processed posts.

@Moriambar yes but are you sure it's looking for tabs at all, if I changed it to remove the or clause and just looked for \ tab it looked like it was finding all \ space

@wilx woff isn't really intended for "desktop" use it's normally just opentype repacked for serving as a web font isn't it?

@DavidCarlisle I'm a bit confused. Let me tell you what I found out. I was able to produce a list with the post that barbara linked last night without looking for tabs, only looking for spaces; you told me that post had tabs; so tabs are reverted to spaces on the db imo

@Moriambar yes you find it if you look for spaces but \ four spaces is rare so almost all the ones with that query are corrupted. Similarly \ tab is rare so if we could find those almost all those would be corrupted. But \ space is pretty common. What @barbarabeeton has been doing is searching for \ space and then checking by eye whether the posting is good or bad, that works but isn't so fast

@DavidCarlisle What I'm trying to say is that there is no \ tab in the db since that gets converted to the four spaces version. Not 100% sure but I have strong feelings about that. We could check with the powers

@Moriambar the tabs are there when it's rendered back to html in the site so I think they must be there but are being normalised in the query (what is the spec for sql like for example? But my plan was to finish the search for the existing four space ones then if necessary go back and ask for more search tools

@Moriambar stopping now, but january 2013 done.

@Moriambar oh the like clause is wrong I think % which I copied from one of the example queries doesn't mean what I guessed it meant tutorialspoint.com/sql/sql-like-clause.htm

@DavidCarlisle I have no idea. But I have hunted down a person who uploaded the raw files on our SharePoint site and asked for TTF/OTF files.

@wilx woff is "Web Open Font Format" so a hint it's intended for web servers:-)

@DavidCarlisle ok great job. I plan to do something tonight, but now I have to stop looking here

@DavidCarlisle the percent sign is the same as the * in other contexts... I will look at the chat in around 7 to 8 hours, it's risky today

@Moriambar yes

@JosephWright thanks for sorting out *.def

@DavidCarlisle Hopefully we've now sorted all of the issues!

@JosephWright I suspect I'm going to get barred from stackexchange data query:-)

@UlrikeFischer saved me from mailing luatex list, I see

@DavidCarlisle You could check by hyperref patch instead github.com/ho-tex/hyperref/issues/25 ;-)

@UlrikeFischer I guess I need to do a pass over the open issues and try to get a hyperref update out after the graphics stuff stable

@DavidCarlisle Don't worry about graphics: I'm on that ;)

@DavidCarlisle I guess next I should look at hyperlinks for L3 ...

@JosephWright want to swap?

@DavidCarlisle Er, 'no'

@DavidCarlisle I've got beamer as well, remember

@JosephWright easy peasy

@DavidCarlisle I invite you to read the beamer source

@DavidCarlisle Guess you are not on the ConTeXt list: interesting discussion about breaking inline math, breqn obviously got mentioned ...

@egreg I've made my latest answer about bugs in drivers CW: do you feel I should do the same for the others?

@JosephWright I was just joking

@egreg Yes but one could argue that in general package authors should be careful abut rep from bugs in their material

@egreg Perhaps a meta post would be appropriate?

@JosephWright I thought that was a perk of the job.

@AlanMunn Fan-didly-tastic! :)

@DavidCarlisle Not for you, as your packages are known to be bug free.

@egreg that is true

@CarLaTeX A friend of mine who lived in the US decided to go to the grocery store. Se asked the cashier if the vegetables she was buying contained preservatives (in general, chemical added to foods to inhibit decomposition). She was supposed to say, "Estes vegetais contêm conservantes?" but she went for the literal translation, "Estes vegetais contêm preservativos?". The problem is, "preservativo" is the fancy, formal name of condom. The cashier bursted laughing!

@PauloCereda LOL

@PauloCereda :)

@PauloCereda There's another English word whose Italian cognate should cause embarrassment. People use it nonetheless, probably not knowing the real meaning it has in Italian.

@PauloCereda And also in Italian it would be embarrassing using “preservativo” for “conservante”.

@egreg I know one in Spanish! Embarazada (ES) means "pregnant", and Embaraçada (PT) mean embarrassing (for female).

@egreg Oh no!

@PauloCereda The same as in Italian

@egreg you don't have to go to other languages than English, because even in English you have words with quite different meanings in different dialects :)

@egreg ooh!

@egreg, @PauloCereda The 'preservative' one is well-known in the UK as the same issue comes up in French

@JosephWright really?

@PauloCereda Relatively, at least: there is a town in France called 'Condom', so this provides a certain amount of amusement in secondary school ...

@JosephWright oh I see

@JosephWright I had the same problem with some French-Swiss girls, many years ago with "Hotel Della Motta". Since the last word in french is a slang word for the intimate parts of a lady :) they were shocked while looking at the hotel (Della Motta was the owner's last name)

hello

@SBM ooh Adele

:)

@PauloCereda Adele?

@SBM Adele has a song named Hello

I could also quote Lionel Richie :)

@PauloCereda oh okay; today was a holiday

@SBM yay, for what?

We have a religious one tomorrow

@PauloCereda for ରଜ (pronounced somewhat like rawjaw)

@SBM cool!

@PauloCereda it's mostly fun for girls and women; for everyone else just yet another day

@SBM What is it, actually?

@PauloCereda just three days for them swinging around and kids playing games and lots of pithas to feast with

@SBM oh

@PauloCereda was it me you were looking for?

@Moriambar LOL

@Moriambar Per non dire delle donne americane che si chiamano Mona. ;-)

@egreg ahahahahahhahahaahahahahah right!

Explanation: in our dialect, “mona” denotes the same intimate part.

And, as adjective, is mostly like “dense, dull”

A screenshot of the web site of our local newspaper: The translation of the subtitle is "Surprised by the police: unidentified people blow up the money machine" (its from an article about an attack on the bank across the street).

@egreg ROFL

@JosephWright @DavidCarlisle My thoughts and prayers with London!

@yo' Is it happened something again? OMG

@CarLaTeX London fire: Six killed as Grenfell Tower engulfed

@yo' Terrifying :'(

@CarLaTeX rofl?

oh so bad :|

@SBM the rolf was referred to a previous message, of course!

@CarLaTeX What does it mean?

@SBM it's a routing algorithm, Routing on Flat Labels

@PauloCereda interesting; what do you use it for?

@PauloCereda you are mean :P

@SBM you analyze the compact structure of a network topology, hence the flat label. It's an algorithm to traverse network nodes and calculate the best route based on a set of premises.

@SBM It means Rolling On the Floor Laughing

@SBM Rescheduling Other Thesis For Laughs @PauloCereda -> of course there was no Thesis in it :D

@Moriambar I gave one possible meaning. Mine is not wrong. :)

@CarLaTeX clearly you meant the algorithm

@PauloCereda :P

@PauloCereda oh I guess it's somewhat like BFS maybe

@PauloCereda how's it going down there?

@SBM it works differently because BFS is exhausting. This is impractical on a real network topology if this one is particulary big.

@Moriambar cold :)

oh

@PauloCereda by cold I'm sure you mean around 30°C

@Moriambar nah, it's 10C

@PauloCereda ok, so kind of cold

@PauloCereda We have 31°C

@Moriambar yep :)

@CarLaTeX ooh

@PauloCereda not much but kind of :)

@Moriambar :D

@DavidCarlisle -- my actual search argument was 'tabular code:"space backslash four spaces" created:2012' and yes, it does pick up things that are already done (don't understand why, unless it works from the database?), and it doesn't pick up the tab items directly, but they're often there somewhere else on a selected page, which is why i prefer to work per page. and this search almost always has the same result if the space before the backslash is omitted.

@barbarabeeton: mail sent, hope you enjoy it. :)

@barbarabeeton Ive learnt some more sql this morning but finding tabs still defeats me, I suspect we should clear the list of four space ones, see what is left in this weekend's dump then go back to the meta question and complain about the complete disregard of the user base ask for some help with searching.

@barbarabeeton the SEDE is updated only weekly

@yo' yes but is the site search box searching SEDE or the live site (I thought it searched the latter)

@yo' -- i know that. but here's how i've been working: activate a search string and see how many items are reported. ("NN results" at top of list.) fix some, then relaunch the search. the number usually changes. in fact, it sometimes changes by itself while i'm working on cleanup. i've assumed that that indicates the search is "live", not from a static database, but maybe there are other reasons?

@DavidCarlisle @barbarabeeton should be the live site, modulo catching/indexing

How long before the next bug in the graphics drivers? Shall I send another update to CTAN? ...

@DavidCarlisle when I get home I think I'll tinker with my second query by excluding the "high valued posters" which I already processed. I think this is the only real filter that can be applied since already filters out the one from the"current per year query" and the accepted answers. What do you think?

@Moriambar but is it finding the tabs at all?

@DavidCarlisle yesterday evening it found the post that @barbarabeeton discovered at our expenses, and which you told me had tabs. But I did nothing to find that since the tabs seem to be hardcoded as four spaces on the database. The only real proof would be to have a post that DEFINITELY has tabs. If that one had, then yes, it finds them and they're equal to spaces

@DavidCarlisle any other tab specific scripts gives only false positives

@Moriambar but how can they be equal to spaces, it doesn't show up with the original search for \<four spaces> and if you look at the text in the edit box it has literal tabs not spaces.

@michal.h21 LuaTeX is way too slow to handle 6k XML files and ~3500 pages of output...

@Moriambar -- i'm pretty sure that one had spaces, not tabs. (but it's been fixed.) i've now changed my search argument to 2011 (everything left on the 2012 list seems to be false positives), and if i come across an example with tabs, i'll leave it and report it here. (but i'm off in an hour to do my exercises in the pool.)

@barbarabeeton thanks

@DavidCarlisle I'll sort that out later.

@barbarabeeton no I reverted the fix it has tabs both in the live site and in last sundays dump

@Moriambar look at tex.stackexchange.com/questions/50352/… and look at the text in the edit box in emacs, the code indentation is with tab and the \ within the body are \<tab> (@barbarabeeton)

@DavidCarlisle I'm at work still and no emacs here

but thanks I'll look into that

@Moriambar that's OK the above comment will stay there forever:-)

@DavidCarlisle I'll come back in about 2 billion years to see if you're right :D

hello.

long time no see!

@nuttyaboutnatty howdy

got a quick non-cricket-related question

trying to write formal-ish business-like letters

not full-on impressive awe-inspiring law firm, but rather plain-vanilla-neat

I'd like to use some form of enumeration

@DavidCarlisle and @Moriambar -- here's another one with tabs: tex.stackexchange.com/q/11401 i've left it untouched. the search i described before (with 2011) is what picked it up. that search also picks up a few false positives; i haven't figured out why.

@barbarabeeton ok thanks. (I haven't read all) but isn't the search about the data dump, thus making the edit immaterial for it?

hazy question:

any pointers to nice enumeration?

@Moriambar -- see this answer by @yo'. he says it's live.

forgot about the above

:)

@DavidCarlisle ^^ you broke LaTeX again

@StevenB.Segletes They are in the current release and have been for some time

@JosephWright So what am i doing wrong? The packages upgraded (in MikTeX) were l3experimental, l3kernel, and l3packages (they revised to 2017-05-31 release). That broke unicode-math, so I upgraded that too (2017-01-30), but it made no difference...

@JosephWright The following code breaks xelatex \documentclass{article}\usepackage{fontspec}\begin{document}Hi mom.\end{document}

@barbarabeeton I think they're not live: otherwise this should have zero for highly rated users data.stackexchange.com/tex/query/683383/…

@StevenB.Segletes My \listfiles output reads

 *File List*
 article.cls    2014/09/29 v1.4h Standard LaTeX document class
  size10.clo    2014/09/29 v1.4h Standard LaTeX file (size option)
fontspec.sty    2017/03/31 v2.6a Font selection for XeLaTeX and LuaLaTeX
   expl3.sty    2017/05/29 L3 programming layer (loader)
expl3-code.tex    2017/05/29 L3 programming layer
l3pdfmode.def    2017/03/18 v L3 Experimental driver: PDF mode
  xparse.sty    2017/05/29 L3 Experimental document command parser
fontspec-luatex.sty    2017/03/31 v2.6a Font selection for XeLaTeX and LuaLaTeX

@StevenB.Segletes when loading both packages

@StevenB.Segletes When there is an 'out of sync' issue with these I know about it really quickly: we do try to avoid it (and it should be increasingly uncommon: they are all tested routinely)

@StevenB.Segletes It sounds like you've got various bits of your system of different 'vintages': this can be an issue with MiKTeX and the 'install on the fly' approach

@JosephWright Roger that. If it proves nettlesome, I may request the helpdesk to reinstall the packages. Thanks for the file list...

@Moriambar (and @DavidCarlisle) -- sigh. not true. i ran a search on my own id, and here's a counterexample that hasn't been totally fixed: tex.stackexchange.com/a/15913/579 . what may be happening with the script is that if more than one instance of the slash + spaces is on an input line, the second one gets skipped. that's what this example looks like -- every other \\ is fixed. and since i know how i work, i know that i clicked on the {} to do the indentation.

@barbarabeeton you mean the elisp script?

@Moriambar as usual emacs causing more trouble than the usual. :) (sorry, I shouldn't interrupt the serious talk...)

@PauloCereda I am not an emacs user, but I think without it we would have never gone this far. Does vim do all of the replacement automatically?

@Moriambar The real replacement thingy is beyond the editor features, it's more of a regex thingy IMHO. :)

@Moriambar David's regex is pretty accurate.

Of course, there might be some corner cases, but we should aim for the majority of occurrences. :)

@PauloCereda I agree. I think that we should have one general query that is proven to work, perhaps given by the powers. Then whatever subquery we use, as long as we are able to find most of it composed by posts to be processed it means that the query is useful. The last posts, as long as they are <300 can be processed later using the general query

@Moriambar Nice plan!

@PauloCereda thanks. I hope the other liquidators see it as you do

I have to admit that it feels strange to be able to contribute while being a total newbie

I like this song. :)

@PauloCereda which movie is it?

@Moriambar Prince of Egypt

@PauloCereda never heard about it. Perhaps I'm too old

@Moriambar nah :)

oh

Hi, I don't have spaces in every table of contents entry... See one of my two questions. Is there somebody here who can help me?

@barbarabeeton not sure what that's a counterexample of, the link goes to an answer that hasn't been edited since 2011?

@DavidCarlisle Just noticed while changing some L3 tests that bringing dvipdfmx.def and xetex.def 'into line' alters how the page size is set in XeTeX. Do you think we should reverse that, as XeTeX has the pdfTeX-like primitives?

@JosephWright What is XeTeX for?

@JordyDeweer it is a unicode version of tex using system fonts

@barbarabeeton that one is bizarre if you look on the site you see \\ at ends of lines, but if you hit edit the \\ go and appear as the usual single backslash follwed by spaces

@DavidCarlisle the XETeX suggestion worked quite well; thank you. Though there's almost no good font for Odia :|

@DavidCarlisle And what are the advantages in comparison with pdfLaTeX or LaTeX?

@JordyDeweer PDFLaTeX generates... well, a PDF. LaTeX generates a DVI file. :)

Oh sorry, I got your question wrong!

This duck is wackier than usual today.

@JordyDeweer hard to say in one line but (pdf)latex is limited to fonts with at most 256 characters which if you are typing arabic or chinese or something isn't so much fun, xetex can use opentype fonts with thousands of characters

@PauloCereda No problem :-)

@DavidCarlisle Oooh, so it's not so interesting for a Dutch (Belgian) user?

@barbarabeeton @Moriambar do you see the same as ^^^^ note I haven't actually edited anything there, the \\ changed to \ in the preview just as I hit edit, and it reverts if I cancel

@JordyDeweer well it's less necessary there but if for example your institution mandates some opentype font set up for Word or other "normal" desktop application, you could either spend a month figuring out how to make a tex compatible version of that font or just use xetex (or luatex) and use the same font directly

@DavidCarlisle Oh, ok, thanks!!!

And which font is normally used in pdfLaTeX? Does that font look good on paper and screen?

@JordyDeweer well whatever font the document specifies. By default it's computer modern (a font designed specifically for tex) but texlive comes with thousands of fonts

@DavidCarlisle Okay, thnx again! And which font do you recommend?

and is it normal my Windows pc takes over three hours to install TeXlive ??

@JordyDeweer that's not really an answerable question. Or at least it hasn't got a useful answer.

@JordyDeweer it used to take a lot longer. It depends a lot on the speed of your network connection

@DavidCarlisle ok and let's hope it is going to finish before 10 o'clock this evening!

@PauloCereda if you hit the edit button in tex.stackexchange.com/a/15913/1090 do you see the same as my image above, that the \\ vanish?

@DavidCarlisle spooky, I do!

@DavidCarlisle Vanished. Very curious.

@PauloCereda I assume the html is a "baked" cached copy from before the corruption, and as soon as you try to edit then you see what's at the back end, but I've not seen it quite like that before (although I haven't looked at the original rendered version in most edits as the sede query link goes straight to the edit page)

@DavidCarlisle Got it, you mean the view is cached, only the edition goes through DB fetching.

@PauloCereda that is certainly true, and why the \\ issue didn't emerge until months after the corruption, as they don't normally regenerate the html views but it would of course be regenerated if that page was edited, but it's not so clear whether the site search is searching the cached html or the live back end, whereas the sede query of course is using the sunday dump of the back end

@DavidCarlisle the rabbit hole seems deeper than we initially thought.

@PauloCereda and the odd thing about tabs is this:

data.stackexchange.com/tex/query/edit/684469?userid=12908

that query shows all spaces (32) but if you look at tex.stackexchange.com/a/50408/1090 and check the text in the edit box you will see it has tabs not spaces

@DavidCarlisle ouch!

@DavidCarlisle: collation issue?

@PauloCereda That was my guess but I tried every collation I could find that didn't give an error and put it everywhere I could put it that didn't give an error (my sql is a bit hit and miss) and nothing made any difference

@DavidCarlisle hmm but once a collection is poisoned, there's no way to revert it IIRC

@PauloCereda have you ever downloaded one of their dumps, what do you actually get?

@DavidCarlisle Never tried to restore them, let me check my sources.

Hi everybody! I’d like to make a minor remark about this question:

But before doing so I must learn how to post code excerpts on this chat… :-)

Can anybody help me?

@GustavoMezzetti If you paste multiple lines you should get 'fixed font' as an option: it's the same 'indent by 4 spaces' as in the main site

@GustavoMezzetti For example:

\documentclass{scrartcl}

\usepackage{showframe}

\begin{document}
   Test
   \vfill
   Test g
\end{document}

@JosephWright So, I don’t need to indent each line by four spaces, do I?

@GustavoMezzetti No, not if you paste in everything with multiple lines: you'll get offered the fixed font button

@GustavoMezzetti if you want to do a single line of code using block formatting you do need to add the spaces by hand

Thank you very much! OK, let’s try it out! I’ve seen that the fix to the \newpage command is implemented like this in the last release of ltoutput.dtx:

\def \newpage {%
  \if@noskipsec
    \ifx \@nodocument\relax
      \leavevmode
      \global \@noskipsecfalse
    \fi
  \fi
  \if@inlabel
    \leavevmode
    \global \@inlabelfalse
  \fi
  \if@nobreak \@nobreakfalse \everypar{}\fi
  \par
%    \end{macrocode}
%    The |\vfil| at the end of the macro before the break penalty will
%    normally result in the page being run short,  even with |\flushbottom|
%    in effect (in contrast to the behavior
%    of |\pagebreak|). However, if there
%    is some explicit stretch on the page, say, a |\vfill|, it has the undesired

@GustavoMezzetti Sounds vaguely right, yes (@DavidCarlisle is the expert)

@DavidCarlisle I've not followed everything, I'm now just trying to see whether my query catches the tabs

@GustavoMezzetti Frank spotted a long standing but and fixed it (hopefully)

@JosephWright Yes. What I’d like to remark is that, for more than 15 years, I’ve been using a slightly different patch, namely

\def \newpage {%
  \if@noskipsec
    \ifx \@nodocument\relax
      \leavevmode
      \global \@noskipsecfalse
    \fi
  \fi
  \if@inlabel
    \leavevmode
    \global \@inlabelfalse
  \fi
  \if@nobreak \@nobreakfalse \everypar{}\fi
  \par
  \@tempdimc\prevdepth
  \ifdim\@tempdimc<\z@
    \@tempdimc\z@
  \fi
  \ifdim\@tempdimc>\maxdepth
    \@tempdimc\maxdepth
  \fi
  \vskip -\@tempdimc \@plus 1fil
  \penalty -\@M}

@GustavoMezzetti 'Talk to Frank'

That is, only one glob of glue instead of two: this should be sligthly more efficient.

Not that it is very important, of course, but…

@GustavoMezzetti Like I say, Frank is the man to talk to here: the key is that we are fixing these things (see LaTeX News)

@JosephWright Well, let’s hope Frank (Mittelbach?) reads this chat!

@GustavoMezzetti He does, but not all of the time: I'll drop him a line

@barbarabeeton @DavidCarlisle this query, when prompted the proper year, catches both your linked posts, thus proving that tabs are caught by the query. Now that query is not refined. You can and should use the previous one while I try and remove some from the latest

@JosephWright Thank you. I repeat, in theory a single \vskip instruction should be preferable to two.

Of course, the conditionals themselves could be made more efficient.

@GustavoMezzetti yes although there might be issues in changing \@tempdimc (but that could be avoided of course)

@DavidCarlisle Thoughts on my XeTeX point earlier?

@JosephWright no I was in a meeting, do I need to scroll up...

@DavidCarlisle I'll post it again :)

@DavidCarlisle Doing some L3 stuff, I notice that the driver changes add an extra \special for XeTeX. That's because I took the dvipdfmx code, which sets the page size that way. For XeTeX, we could switch back to using the pdfTeX-like primitives: I suspect this might be 'sensible'. Thoughts?

@JosephWright It's probably better to use the pdftex primitives (as users may set them directly and then you have to worry about what it means if they are in conflict with teh \special

@DavidCarlisle Yes, that's what I'm thinking: I've got another release to do anyway so I'll alter that too

@barbarabeeton @DavidCarlisle this query should be the last to be dealt with after the other we're going through (please remember they complement each other) data.stackexchange.com/tex/query/684580/…

@Moriambar I'm going to have to go offline in a bit but I'm confused, because isn't that just (apart from the extra filters) the original query of replace space by X then look for \XXXX ?

@DavidCarlisle yes, that proves that it takes into account tabs if your previous assumptions were correct, ie that the two posts have tabs. To say it in technical terms it was not the fact that there were tabs that made them not pop up, but their null accepted answer value

anyway the latest query gives around 1800 to be added to the current one, but I feel that many have already been done. Nevertheless I have the 2013 july month from the prev query to go through

@Moriambar I am sure that ones with just a tab (and not three spaces as well) come out as just \ so \X not \XXXX so are not found even by the original query with no extra filters, but I need to check. That is tab is appearing as space but searching for \ gives too many false positives

@DavidCarlisle Indeed, there should have been a reason for which I chose \@tempdimc, althogh I cannot recollect it now—I wrote the patch back in 2000. In any case, I have been testing it since, without experiencing problems… :-)

@DavidCarlisle fine, I think we can deal with those whenever we finish the ~2000 we have left, in order to have a smaller pool of data to look onto

@GustavoMezzetti put it on ctan and wait ten minutes to find someone setting \@tempdimc the line before a \newpage ask @JosephWright aboy these things:-)

@JosephWright got to go, clarinet to listen to....

@DavidCarlisle nice

@DavidCarlisle Good plan

@DavidCarlisle :)

@DavidCarlisle OK, use \begingroup ... \endgroup then! :-)

Anyway if @barbarabeeton or @DavidCarlisle finds one of the tab as \X please do not edit it, but stop and tell me in order to fix the script, or try at least

Well, that’s all I had to say: goodbye to everybody, have a pleasant evening!

goodbye

July done

Also:

@PauloCereda Ducks are always in trouble...

@TeXnician we are! :)